The LLM is reading the bytes of the file, not looking at a picture of its rendering. File metadata exists as well, and change history. Tons of places to hide text.
Even if you processed it via a screenshot, image files are processed byte by byte as well and can contain textual metadata.
TLDR: “we’re going to try to get money from LLM providers for access to our back catalog without getting permission from the authors or providing them with any share of the revenue”.
China, Russia, and plenty of other countries have all been actively and successfully destabilizing the US for well over a decade now. To the point that your argument about who is less a threat to Americans depends heavily on the skin color, religion, and ethnicity of those Americans.
Ironic to oppose giving AI tech to “authoritarian governments” while approvingly quoting the authoritarian-wannabe government of the US and framing that government as the good guys.
The “contents” represent the majority of the cost and meaningful functionality of what we call a “datacenter”. Those contents will not last for “real estate” debt timelines.
Right but their payback period is insanely short - a $5M GB200 NVL72 cluster is expected to generate $75M in revenue over 3 years for inference providers. That's a 3 month payback period.
AND - they're operating well past their estimated service lifetime.
Chances they're planning on replacing personal, local compute with time-sharing on data center hardware that's too outmoded for AI...? You know, since they sunk the consumer component market for the next half-decade.
“In order to serve our customers better, we are phasing out the ‘Pro’ and ‘Plus’ plans, which will be replaced with new ‘Premium’ and ‘Extra’ plans. In addition, our ‘Business’ plans are now being replaced by improved ‘Corporate’ plans.”
The idea that we can’t get enough electrical power on earth because of supply constraints, but that we can solve that by launching the datacenters and their entire support infrastructure into space makes no sense unless you’re a huckster trying to pump up your space launch business ahead of an IPO. There’s no way we are more constrained on land-based solar+battery equipment than we would be on space-worthy solar+battery+space-cooling equipment plus sufficient launch infrastructure and resources. Meanwhile, our government is paying companies to stop building wind power and bombing sources of natural gas.
sheesh i didn't say a datacenter-in-space was a good/feasibnle idea. The point i was making is I feel that building power plants to power a datacenter is turning out to be harder than software people realize and multi-year (or worse) delays likely the result.
The space required to install enough solar panels, especially in the places these things are going (like Ohio, where it's NOT ideal for solar generation), would far, far outsize the data center itself by a significant magnitude.
It takes 2.5 million solar panels to generate 1GW of electricity assuming 400W per panel and direct sunlight. 2.5M times 18 sq ft per panel is about 1000 acres or ~1.55 sq mi.
That would work for a 1 GW data center, during the day, without clouds, during the summer.
The US gov't owns at least 50% of every state in the west. Or they could just build in the middle of bum f*ck Texas and get fossil and renewables. Too bad Texas' grid is so piss poor
If only data centers had some kind of backup systems to prevent power interruptions, like inverters with batteries attached to them. They could call it an Uninterruptible Power Supply and then charge the batteries when other power is available and run on them when it isn't.
Anything using 1GW of power is going to be "enormously expensive". The all-in cost of a 1GW data center is on the order of $50 billion. A billion dollars in storage batteries is ~2% of the cost, which is completely reasonable.
~1600 GWh of lithium batteries are produced each year and hundreds of GWh/year of production capacity have been added in each of the last few years, so ~12 GWh isn't going to upend the market.
If grid storage batteries are economical in the general case then why wouldn't they be doubly so in a place where the inverters are a sunk cost and you're saving the transmission costs because the load is at the same site as the batteries?
The competition isn't the cost of generators, it's the cost of grid power, or the cost of operating gas generators every day.
They aren't economical in the general sense. They are economical to provide frequency stabilization, grid efficiency, and maybe a few hours of demand shifting. They're not yet economical to store solar power during the day and discharge it at night. We are getting closer but I don't think we are quite yet there.
As far as I can work out, the cost is competitive as long as you're using the capacity every day, especially when it lets you avoid paying transmission costs. The thing that doesn't work is to have enough battery storage to handle extended periods of low generation, because for that you need dramatically more storage capacity while only using the surplus on rare occasions. But a data center could use the grid or backup generators on those rare occasions while still using batteries overnight every night.
Space DC make sense when local governments and communities don't want DC built on Earth. Space DC is an engineering and finance problem, not a problem in legal/permission, real estate, power generation, grid delivery, environmental, etc.
>a less hard problem to solve than bribing a few local governments
Wouldn't it be best if the data centers were just... good neighbors? Like everyone talking about illegal bribes and impossible moonshots when the answer is just don't suck?
We are well past the point of these companies caring about their reputations. People are already losing their homes over this. Here is Penguinz0's commentary on a news report of GA Power eminent domaining someone's home to run a power line through it to power a data center.
The general problem is that we've made construction suck for a while now and people are suddenly paying attention to it but only in the case of data centers.
You don't get caught paying bribes if government officials aren't accepting bribes. The problem is the governments have been operating in a mode where blocking everything is the default and paying bribes is the expected way for things to get approved.
The issue with this is that datacenters are inherent "suck" machines. It takes a lot more effort and cost for them to not suck then for them to suck, so "sucking" is part of their core design.
Up till recently, they were a rare enough object that distressed communities still saw the glitz and glam that they are sold with as valuable. Now that brutalist assholes are trying to jam them in every nook and cranny of the earth, word is getting out and people aren't buying it anymore.
There have been some attempts to try floating or water immersed data centers that mitigate or at least hide some of their suckiness, but these were hard and costly, so the incentives weren't there to do them anyway. Maybe now those ideas will resurface and improve, but I expect "bribe, con or sue rural community to destroy their quality of life" is still cheaper and easier.
You still need to build the power infrastructure to the rural areas. Build the infrastructure for the people, like roads, sewage, water, buildings, etc. Navigate the regulation and permit processes.
Sending to LEO costs around $2,700 usd and 9kWh of fuel per kilogram.
If you manage to send a data center with high end processing units, large array of solar panels, even larger liquid cooling and tick shielding against solar rays you’d have spent probably 100x to 1000x the cost of an equivalent earth data center even without considering hardware deprecation.
You’re better off sinking your data center deep in the sea and even that is probably not happening at large scales.
I tried to find what the worst option is other than putting them in space and you can do a lot! Trucking diesel to the middle of the Sahara is a decent option in comparison to space.
How do the laws of physics allow it? How do you disperse the heat? How do you protect your compute from cosmic rays? How do you get enough solar up there without catastrophic damage from micrometeorites?
I think the company that has managed to launch and currently operates ~10k satellites can figure those out. Still remains to be seen if the economics make sense, but with increased "anti" stuff happening all around (some of which surely is manufactured) who knows? 5 - 10 years from now things might change.
Oh, well, if the CEO of a company who only exists to try to convince people it is possible says it's possible on a podcast hosted by a VC firm who also wants to convince people it is possible so they can make their money back then it must be!
I've always wondered who is gullible enough to even bother watching this kind of garbage, thank you for answering that question.
Space data centers are an intrinsically cool idea, and honestly I hope we see at least one for the sheer marvel of such a project. But to be cost efficient, it has to be cheaper than rural US, cheaper than buying some island, cheaper than setting it up on a converted freighter ship and tanking the fuel in, etc.
BUT once Earth has a fully paid off space elevator the unit economics will make sense, or if we have orbital space factories and don't have any lift costs, etc. It's all possible, but at a 50 year+ timescale and immense investment. All of which I hope for because I'd like to see more space development and exploration.
On Earth it has to be fossil fuels. With solar you can't do it. I haven't done the calculations myself so, full disclosure, AI work ahead. But I asked GPT 5.6 how big a solar and battery farm would have to be to power a one gigawatt datacenter such that it could ride out a few cloudy days in a row (this is far below the reliability bar).
Its answer is you'd need an "absurdly large battery" and a solar farm of over 100 square kilometers, quite possibly up to over 200 square kilometers. And it would need to be built in the desert.
Also, normal utility scale LiON battery sites are only meant to provide stabilization services for hours, not days.
The problem is you have to overspec the solar farm quite a lot because it must not only be able to power a one gigawatt facility, but also generate more than that so it can recharge the battery packs during the day too.
So this just doesn't work. You think datacenter permitting is bad when it only requires some on-site gas turbines? Imagine the difficulty of getting a 200 square kilometer solar farm over the finishing line.
No, it's quite possible Elon is going to win this one. Regulation is expensive! He can just fire a constant stream of satellites into space. The constellation needed to provide a gigawatt of inferencing would be about the same size as today's Starlink system, well within the bounds of what he's proven can be done.
The threat to space based datacenters isn't really the physics of it. It's the risk of a bubble pop that temporarily craters demand and/or an overbuild of terrestrial capacity. Outside of the software industry we don't see huge new blocks of inferencing demand, and OpenAI at least doesn't seem capacity constrained at all right now.
Why restrict yourself to 100% solar? Solar augmented by grid is pretty great.
Also, you very conspicuously avoided talking about the difficulty/impact of a 1GW datacenter in space. One might suspect your motives aren't as impartial as you seem to think.
The whole problem is difficulty getting grid connections on Earth. If you can just hook up to the grid then why not just use it for everything.
I don't have any motives, what are you talking about? I don't own SpaceX stock if that's what you're thinking.
What's the difficulty of a 1GW datacenter in space? The number of satellites it requires seems to be within the realm of what SpaceX has already achieved. If the calculations are wrong, by all means say so, I haven't checked them.
> What's the difficulty of a 1GW datacenter in space?
Simple back of the envelope calculations?
Let's assume your numbers are correct and that panels in space can generate 10x the power per square meter per 24 hour period (due to efficiency, lack of night, and lack of atmosphere/weather). That means 10 KM^2 are needed for one such data center. The largest we've built in space is ~3,000 M^2 on the ISS. So one of these DCs in space will require an array ~3,000 times the size of the largest we've built before. Or, it's going to require ~3,000 satellites, each with an array the size of the largest we've ever deployed.
The largest Starlink satellites currently deployed generate less than 30 KW each. We'd need more than 30,000 of those to generate this much power. Starlink has launched ~12,500 to date, with ~10,000 still functioning. So 3x the size of the functioning fleet that's taken 8+ years to deploy. Currently, we are deploying well under 5,000 per year. Let's assume we can come up with enough spare capacity to launch 5,000 per year. That's 6 years to deploy the first 1GW DC. By the time we get to 1GW, the average GPU is 3 years old. So we need to continue to deploy 5,000 satellites per year to maintain 1GW of GPUs that are, on average, 3 years behind current generation GPU designs.
This is all before we get to the economics of launching, the cost of the satellites themselves, heat dissipation, radiation hardening, hardware failure rates (~20% of deployed Starlink satellites are no longer functioning), etc.
The whole thing does assume Starship works, so at that point satellites could be launched much faster.
I assume Starmind would use Tesla designed chips, which removes Nvidia's margin and so the cost of the chips is not necessarily so high.
Rad hardening isn't important at low altitudes, and inferencing can smoothly recover from glitched calculations/lockups.
I don't know if it makes economic sense or ever will, but technologically it does seem possible. And if you can't build on Earth for political reasons, then space can win by default.
In terms of what we decide to build, impossible vs. impractical is not a particularly interesting distinction (that distinction can give rise to fun science fiction). Practicality is at the core of good engineering, and possibility is just the first of many steps in evaluating practicality.
If we're going with a distributed approach, there are other ideas that are more practical. For example, send 1 KW boxes to a million people (< 1% of American homes, but there's no reason to limit this to the US). Very roughly, these are beefed-up gaming consoles - you could even start with existing gaming consoles. That box plugs into electricity and ethernet. Pay each of those people $1,000 a month to leave their box running 24/7. For people with a lot of solar at home, it could be a nice income stream. At $1B per year for payouts, it's far cheaper than doing it in orbit. And it doesn't rely on promised but as-yet undelivered technologies. But data centers are probably more practical.
1. Running modern models requires a whole rack of extremely loud equipment that can easily consume 130kw. That's not a load you can or should run at home.
2. It would require a very strong end-to-end implementation of confidential computing to keep the prompts encrypted. If there is any flaw in that infrastructure at all, like someone discovers a way to physically tamper with the hardware to read the GPU memory, then the entire network is a writeoff and would have to be replaced. That's a massive risk. Confidential computing can work in cases where hardware replacements aren't an issue or where the vendor has a long history of successfully defending the platform. But Nvidia is the only GPU vendor with these capabilities and their CC is untrusted.
Putting the workloads in space means noise and data privacy aren't an issue, assuming you trust SpaceX to not peek at your prompts on their way to the cluster.
> The whole problem is difficulty getting grid connections on Earth. If you can just hook up to the grid then why not just use it for everything.
Because you already need inverters and batteries for backup power and then using solar + batteries for ordinary operations and the grid for backup is cheaper than the other way around.
Which is also why you may not even need the grid connection. Suppose you have diesel generators to use in case you have some kind of a fault in the solar system, the same as was traditionally done in case of a failure of a grid connection, and then use those for the one week every two years when solar generation is too low.
(1) these are not ordinary radiators because ordinary radiators aren't gold plated and don't unfold dramatically from compact shipping sizes
(2) It's absolutely loads of radiators
100 square kilometers sounds like a big area. And it is, in continental Europe or in an urban area. In Montana, it's enough space for a few hundred cows. In desert regions, it's worth even less. You can buy 100 square kilometres of unserviced desert land for ~$10M.
Even better, get rid of the mandate for ethanol and cover a portion of the ex corn fields with solar panels with native plants under and around them. A small portion of the land that is near existing transmission lines is needed to replace combustion vehicles with electric vehicles. The bulk of the rest of this land could be available for other uses, including data centers.
This gives: sustained income for landowners, CO2 reduction, power for new uses, and partially restored ecosystems that have been greatly harmed by chemical laden monocultures.
Twain seems the obvious choice, so yeah I’m disappointed they didn’t even mention him. Given that they invited submissions, I’d have loved to see a bigger variety, rather than two Frosts and two “no one”s. Those are both good choices, but not super interesting to read two of when other choices are left out.
Even if you processed it via a screenshot, image files are processed byte by byte as well and can contain textual metadata.
reply